Since the frequency measurement on my Rigol bench multimeter is maxed at 3MHz I was planning for some time to get a dedicated frequency counter. My first choice was one of the Philips PM6669 or PM6666 counters. They can be obtained for a reasonable price and are also compact in dimensions as that is one of my biggest concerns considering the lack of desk space.
So as soon as I saw a listing for Philips PM6666 with PM9608B (1.3GHz option) for a reasonable price, I had to get it.
Seller was extremely forthcoming and with the instrument he sent a ton of documentation (datasheet, manual, service manual etc.). He also told me that he had to replace the input filter since the original one burned. Thank you CCO.
Unit itself was working and was in great condition for the instrument made in 1992/1993.
Non the less, since I plan to use it for years to come I decided to do a capacitor refresh.
Service manual contains a BOM and I used it to make an order for new capacitors.
Only capacitors with chemistry prone to aging were replaced:
This is a picture of a PCB after a recap:
One thing I was not satisfied with before and after a recap was temperature stability.
Every small temperature change was immediately visible as a change in measured frequency.
That issue bothered me a lot so I was determined to find a solution. Thankfully I wasn’t the first person facing that problem.
A user “miedema” on Netherlands forum designed a complete solution and manufactured a PCB. Forum post can be found on this link: https://www.circuitsonline.net/forum/view/message/1776614
Although the post is now four years old I contacted him and asked if he has any spare boards left to sell.
He was most kind and replied affirmatively. I quickly accepted more than fair offer for a PCB and an VCTCXO. And on top of that he added an extra VCTCXO and two 50PPM resistors used for a voltage divider together with all the documentation, parts list and datasheets. Thank you Gertjan.
With the hardest part done by Gertjan only the easy one remained. I ordered the necessary parts and proceeded with soldering.
That part was no simple feat because of the big ground surfaces that were acting quite efficient at sinking heat from my soldering station. That and the warnings about how any excess heat may damage VCTCXO or at least increase its initial drift. So I was extra careful to be quick and not to apply any unnecessary heat.
The only part I could not obtain was a voltage reference IC: REF3133AIDBZT. So I had to use its closest replacement: REF3233AIDBVT.
Since it is not direct pin to pin compatible I had to do a bodge on a PCB but it works flawlessly.
Photos of Philips Fluke 1.3GHz upgrade option PM9608B are quite scarce on internet so I added few of them together with original oscillator in hope that someone finds them useful.